The present invention relates to novel active fragments of basic fibroblast growth factors (bFGF). The fibroblast growth factors (FGF) are multifuntional polypeptide mitogens which exhibit broad target-cell specificity (1). In the course of study of these factors, a number have been identified on the basis of the ability of extracts from various tissues, such as brain, pituitary and hypothalamus, to stimulate the mitosis of cultured cells. Numerous shorthand names have been applied to active factors in these extracts, including epidermal growth factor, platelet-derived growth factor, nerve growth factor, hematopoietic growth factor and fibroblast growth factor.
Fibroblast growth factor (FGF) was first described by Gospodarowicz in 1974 (2) as derived from bovine brain or pituitary tissue which was mitogenic for fibroblasts and endothelial cells. It was later noted that the primary mitogen from brain was different from that isolated from pituitary. These two factors were named acidic and basic FGF, respectively, because they had similar if not identical biological activities but differed in their isolectric points. Acidic and basic fibroblast growth factors (recently reviewed by Burgess, W. H. and Maciag (3) appear to be normal members of a family of heparin-binding growth factors that influence the general proliferation capacity of a majority of mesoderm and neuroectoderm-derived cells (4), including endothelial cells, smooth muscle cells, adrenal cortex cells, prostatic and retina epithelial cells, oligodendrocytes, astrocytes, chrondocytes, myoblasts and osteoblasts (Burgess and Maciag, cited above at Page 584) (3). Although human melanocytes respond to the micogenic influences of basic fibroblast growth factor but not acidic FGF, most avian and mammalian cell types respond to both polypeptides (ibid.) (3).
In addition to eliciting a mitogenic response that stimulates cell growth, fibroblast growth factors can stimulate a large number of cell types to respond in a non-mitogenic manner. These activities include promotion of cell migration into wound areas (chemotaxis), initiation of new blood vessel formulation (angiogenesis), modulation of nerve regeneration (neurotropism), and stimulation or suppression of specific cellular protein expression, extracellular matrix production and cell survival important in the healing process (Burgess and Maciag, cited above, pages 584 to 588) (3).
These properties, together with cell growth promoting action, provide a basis for using fibroblast growth factors in therapeutic approaches to accelerate wound healing and in prevention and therapeutic applications for thrombosis, artheriosclerosis, and the like. Thus, fibroblast growth factors have been suggested to promote the healing of tissue subjected to trauma (5), to minimize myocardium damage in heart disease and surgery (6 and 7), and to increase neuronal survival and neurite extension (8).
Complementary DNA clones encoding human acidic and human and bovine basic fibroblast growth factors have been isolated and sequenced, and the predicated amino acid sequences derived from the complementary DNAs agree with the structures determined by protein sequence analysis (summarized by Burgess and Maciag, cited above, at pages 580-581) (3). The data predict acidic fibroblast growth factor (hereafter referred to as aFGF) to have 155 amino acids (ibid) (3). The gene for basic fibroblast growth factor (hereafter referred to as bFGF) also codes for a 155 residue protein. For both aFGF and bFGF N-terminally truncated forms exist that exhibit full biologic acitivity, including the 146-amino acid bFGF orginally isolated and sequenced (9) and a 131-amino acid form. Analysis of the structures demonstrates a 55% identity between aFGF and bFGF (Burgess and Maciag, cited above at page 581) (3).
Basic fibroblast growth factor may be extracted from mammalian tissue, but this requires several steps even when heparin-linked affinity chromatography is employed (U.S. Pat. Nos. 4,785,079 and 4,902,782 to Gospodarowicz, et al.) (10 and 11), and the 146-amino acid species is generally obtained if extraction is done in the absence of protease inhibitors (ibid., column 9, lines 29 to 32). Bovine and human basic fibroblast growth factor cDNA have been expressed in E. coli (12 and 13) and S. cervisiae (36). However, reported yields of product are low (15), and recombinant factors exhibit a marked tendency to undergo thiol-disulfide interchanges promoted by free thiol groups in the protein that result in the formation of disulfide scrambled species (12).
A number of basic fibroblast growth factor analogues have been suggested. Muteins of bFGF having amino or carboxyl terminal amino acids deleted, amino acids added, cysteine substituted with a neutral amino acid such as serine, or aspartic acid, arginine, glycine, serine, or valine substituted with other acids have been suggested to have enhanced stability (16). The muteins comprise two or three additions, deletions or substitutions, with substitution of serine for cysteine the most preferred substitution (16). Arakawa and Fox (17) suggested replacing at least one, and more preferably two, of the cysteines found in natural bFGF with a differnt amino acid residue to yield a more stable analogue (page 4, lines 44 to 47); serine was illustrated in the Examples (page 13, lines 22 to 23). Similarly, recombinant aFGFs having extraneous bond-forming cysteine replaced with serine, and oxidation-prone cysteine, methionine and trypotophan replaced with alanine, valine, leucine or isoleucine, to yield factors having enhanced or improved biological activity have also been suggested (18).
A bFGF mutein lacking 7 to 46 amino acids from the carboxyl terminus and, optionally, having amino acid replacements was suggested to have improved stability while retaining acitivity in Eur. Pat. Ap. Pub. No. 326,907 to Seno, et al. (page 2, line 50 to page 3, line 4) (19). Fiddes, et al, (Eur. Pat. Ap. Pub. No. 298723) (20) suggested replacing basic or positively charged residues in the heparin binding domain encompassing residues 128 to 138 with neutral or negatively charged amino acids to produce forms of FGF having reduced heparin binding ability and enhanced potency (page 5, line 45, and page 5, line 54 to page 6, line 16). Bergonzoni, et al.(21), suggested six analogues: 1) M1-bFGF, lacking residues 27 to 32; M2-bFGF, lacking residues 54 to 58; M3-bFGF, lacking residues 70 to 75; M4-bFGF, lacking residues 78 to 83; M5-bFGF, lacking residues 110 to 120; M5a-bFGF, having the position 128 lysine and the position 129 arginine replaced with glutamine residues; and M6b-bFGF, having the positions 119 and 128 lysines and the positions 118 and 129 arginines replaced by glutamine residues.
Fortunately, the affinity for heparin also provides a selective method for isolation and purification of the two forms of FGF, acidic and basic (22). FGFs are structurally labile but can be protected from inactivation by heat or low pH by association with heparin (23). Heparin-FGF complexes also are highly resistant to proteolytic degradation (24). Heparin, through a mechanism most probably related to enhanced stability of the fibroblast growth factor can potentiate the biologic properties of both acidic and basic FGF (23 and 25). FGFs lack a classical signal peptide sequence which can direct secretion of the FGF into extracellular space (26). However, considerable quantities of bFGF have been detected in and isolated from the extracellular matrix (ECM) both in vitro (27) and in vivo (28). This, therefore, raises the question of how secretion of FGF occurs and suggests the possible use of a carrier protein or proteoglycan. ECM-associated bFGF is bound to heparin sulfate (HS) proteoglycans (29 and 31) and is most likely released in a controlled manner a as FGF-HS complex (31). Thus, a current hypothesis (1) regarding the mechanism for regulation of the mitogenic activity of FGF is that FGF is sequestered in the ECM as a biologically inactive FGF-heparin sulfate proteoglycan complex. Then, ECM-bound FGF may be mobilized by specific cellular signals that activate, for example, matrix-degrading enzyme systems such as the plasminogen activation cascade (31) and heparan sulfate-specific endo-beta-D-glucuronidases (32 and 33).
As such, it seems that locating the functional domains on bFGF should involve reviewing this structure function interaction with heparin and the receptor for FGF. In fact, structure function studies with synthetic peptide fragments of bFGF suggest the existence of two functional domains, corresponding to residues 33-77 and 112-155 (The numbering for bFGF adopted here is for the 155 amino acid form as described in (26), and refers to the methionine initiation codon as position 1. Using this system the N-terminal prolyl residue of the sequenced 146 amino acid tissue derived form of bFGF corresponds to codon position 10 and is thus identified as bFGF (10-155)). Heparin and receptor-binding domains were identified by the ability of synthetic peptides related to FGF sequences to compete with bFGF for its receptor, to bind to radiolabeled heparin and to modulate the mitogenic response of FGF (34). From one of these regions containing a functional domain, 112-155, an active core decapeptide (115-124) was produced. However, the peptide (115-124) had decreased affinity for heparin and biologic potency by 10-and 100-fold, respectively in comparison to peptide 112-155 (34).
Seno et al (35) provided another series of fragments to study the mitogenic and heparin-binding properties of a series of C-terminally truncated bFGF's (10-155) produced in E. coli. As the degree of C-terminal truncation exceeded 6 residues, affinity for heparin was markedly decreased. In the same study, two N-terminally truncated forms of bFGF, 23-155 and 50-155, were also studied. When compared to the native bFGF (10-155), bFGF (23-155) and (50-155) showed no changes in affinity for heparin but exhibited about a 2-and 50-fold decrease in mitogenic activity, respectively.
The present invention takes advantage of the specific significant interaction between heparin and bFGF to more closely define the structural domains of the growth factor that interact with heparin and the receptor for bFGF to yield a mitogenic response. The human bFGF mutant glu.sup.3,5, ser.sup.78,96 bFGF is used for these studies for the following reasons: (a) bFGF and the mutants are equipotent; (b) the mutations of residues 3 and 5 afford considerable higher expression than the parent protein in the expression system of this invention; and (c) the mutations of cysteines at residues 78 and 96 to serines eliminates disulfide scrambling-related stability problems associated with recombinant preparations of the parent bFGF, particularly when expressed in high yield.
Proteolytic digestion of heparin sepharose-bound glu.sup.3,5 ser.sup.78,96 hbFGF(1-155) gives two peptide fragments that are eluted from heparin sepharose under the same conditions used to elute intact bFGF. Thus, these bFGF peptide products are heparin-binding fragments (HBF), since they are protected from proteolytic degradation by virtue of their specific interaction with heparin. The 2 peptide fragments are labeled HBF-1 (bFGF 27-69) and HBF-2 (ser.sup.78,96 bFGF(70-155)). On heparin affinity HPLC the mixture of HBF-1 and HBF-2 is not resolved, and the two fragments coelute with a retention time identical to that of intact bFGF. It is assumed, therefore, that both fragments possess equal affinities for heparin; although the possibility that they are covalently linked or non-covalently associated to form a complex, also exists. Interestingly, HBF-1 and HBF-2 each contain a single cysteine residue corresponding to positions 34 and 101 in glu.sup.3,5, ser.sup.78,96 hbFGF(1-155) which are thought to be disulfide linked in the native structure (35). HBF-1 and HBF-2 are, however, efficiently resolved by RPLC in the absence of a thiol reducing agent. The fact that N-terminal sequence analysis of RPLC-purified HBFs indicates no secondary sequences argues against HBF-1 and HBF-2 being covalently linked.
Thus, HBF-1 is subjected to S-pyridylethylation (37) under non-reducing conditions, followed by N-terminal sequence analysis. This procedure gives a quantitative reaction with vinylpyridine and release of phenylhydantoin S-pyridylethyl cysteine derivative on the 8th sequencer cycle, whereas a similiary treated control, somatostatin-28 (Bachem Bioscience), in which cysteines 17 and 28 are disulfide-linked, gives no detectable phenylhydantoin- cysteine derivative at sequencer cycle 17. Thus, the data support the presence of a free sulfhydryl group in HBF-1, and by analogy, in HBF-2, as well. This implies that cysteines 34 and 101 may not be stably disulfide-linked in glu.sup.3,5,ser.sup.76,96 -bFGF.
Alternatively, the possiblity exists of non-covalent interactions between the 2 peptides. Indirect evidence for an association between HBF-1 and -2 comes from the failure of cation exchange chromatography, despite the marked differences in net charge of the peptides and lack of disulfide linkage, to resolve the 2 fragments. Also, their behavior on both heparin and cation exchange chromatographies is indistinguishable from that of bFGF. This may imply that in the native state or when bound to heparin, regions of bFGF contained in the sequences 27-69 and 70-155 interact to yield a unique 3-dimensional structure that is required for high affinity binding to heparin. Additional support for this comes from the observation that RPLC-purified bFGF, HBF-1 and HBF-2 do not retain their affinity for heparin, even after removal of HPLC solvents, but RPLC-purified bFGF and HBF-2 exhibit biologic activities.
Baird et at (34) using synthetic bFGF fragments identified 2 peptide regions, residues 33-77 and 109-129 in bFGF, that bind to heparin and exhibit weak partial against antagonist activities in biological assays. HBF-2 (ser.sup.78,96 bFGF(70-155)) contains the bFGF sequence 109-129. A comparison of the potencies of FGF sequences relative to intact bFGF (Table 2) shows that the activity of HBF-2 is at least 10.sup.3 -10.sup.4 fold greater than the most active synthetic peptide known (bFGF (112-155)) (34). Recently, Seno et al (35) expressed and examined the properties of a series of C-terminally truncated versions of bFGF. These studies conclude that essential elements for receptor binding are contained in the sequence 50-109 and for heparin binding in the sequence 110-150; however, other interpretations are possible. Seno et al., also described an N-terminally truncated form of bFGF (50-155) which retains full affinity for heparin and exhibits about 2% the mitogenic activity of bFGF (10-155). Since bFGF (50-155) and HBF-2 (70-155) seem equipotent (Table 2), the FGF sequence 50-69 can thus be eliminated as contributing significantly to receptor recognition and the mitogenic response.
The present invention provides FGF fragments obtained upon proteolytic degradation of bFGF bound to heparin or heparin sepharose and include fragments which immobolize heparin sepharose analogues or heparin analogues, including ones with mitogenic activity. This invention also encompasses analogues having the cysteine residues at positions 78 and 96 replaced with other amino acids, such as, for example alanine, glycine, arginine, tryptophan, lysine, aspartic acid, glutamic acid, asparagine, glutamine, histidine, isoleucine, leucine, valine, phenylalanine, tyrosine, methionine, serine, threonine or proline. More over, the bFGF fragments of the present invention are not species specific, and include instance, bovine, ovine, porcine and others that share similar sequences homology with human bFGF.
The novel bFGF fragments of the present invention also may be prepared by recombinant protein synthesis involving preparation of DNA encoding the bFGF fragments, insertion of the DNA into a vector, expression of the vector in host cells, and isolation of the recombinant bFGF fragments thereby produced.
Because of the degeneracy of the genetic code, a variety of codon change combinations can be selected to form DNA that encodes the bFGF fragments of the present invention, so that any nucleotide deletion(s), addition(s), or point mutation(s) that result in a DNA encoding the bFGF fragments herein are encompassed by this invention. Since certain codons are more efficient for polypeptide expression in certain types of organisms, the selection of gene alterations to yield DNA material that codes for the bFGF fragments of the present invention are preferably those that yield the most efficient expression in the type of organism which is to servive as the host of the recombinant vector. Altered codon selection may also depend upon vector construction considerations.
DNA starting material which can be altered to form the DNA of the present invention may be natural (isolated from tissue), recombinant or synthetic. Thus, DNA starting material may be isolated from tissue or tissue culture, constructed from oligonucleotides using conventional methods, obtained commercially, or prepared by isolating RNA coding for bFGF from fibroblasts, using this RNA to synthesize single-stranded cDNA which can be used as a template to synthesize the corresponding double stranded DNA. As pointed out herein, proteolytic degradation of native bFGF bound to heparin or immoblized heparin can generate the fragments, as well.
Also encompassed are DNA sequences homologous or closely related to complementary DNA described herein, namely DNA sequences which hybridize, particularly under stringent conditions, to the cDNA described herein and RNA corresponding thereto.
DNA encoding the bFGF fragments of the present invention, or RNA corresponding thereto, can then be inserted into a vector, e.g., a pBR, pUC, pUB or pET series plasmid, and the recombinant vector used to transform a microbial host organisms. Host organisms may be bacterial (e.g., E. coli or B. subtilis, yeast (e.g., S. cervisiae) or mammalian (e.g., mouse fibroblast). Culture of host organisms stably transformed or transfected with such vectors under conditions facilitative or large scale expression of the exogenous, vector-borne DNA or RNA sequences and isolation of the desired polypeptides from the growth medium, cellular lysates, or cellular membrane fractions yields the desired products.